Create a copy-paste prompt pack for a researcher working on the project below. [RESEARCH QUESTION] [FIELD AND STUDY CONTEXT] [AVAILABLE SOURCES OR DATA] [METHODS AND CONSTRAINTS] [OUTPUT OR PUBLICATION GOAL] Produce 8 standalone prompts in labeled blocks: research-question refinement, search strategy, literature-screening rubric, evidence table, study or analysis plan, qualitative-codebook draft, results-to-claims audit, and abstract or manuscript-outline draft. Each prompt must state exactly what I should paste and request a structured output. The search strategy must give concept blocks, synonyms, inclusion and exclusion criteria, and database-ready Boolean logic, while warning that I must run and document the search myself. The evidence-table prompt must require citation details, study design, population, measures, limitations, and direct support for each claim. The methods prompt must distinguish exploratory from confirmatory work, identify confounders or threats to validity where relevant, and flag ethics, consent, privacy, preregistration, or statistical-review questions for qualified oversight. Do not invent studies, DOIs, quotations, results, citations, data, code execution, or methodological approvals. Never turn a summary into evidence of causality beyond the supplied studies. Mark missing information and contradictory findings clearly. Before answering, check that every prompt preserves the source trail and separates finding, interpretation, and limitation. Ask up to 3 clarifying questions only if a required input is missing.
Fill in
| Placeholder | What to enter | Example |
|---|---|---|
| [RESEARCH QUESTION] | State the question, population, phenomena, and outcome you want to study. | Among first-year community-college students, how is commute time associated with course attendance? |
| [FIELD AND STUDY CONTEXT] | Describe the discipline, theoretical frame, setting, and intended study type. | Higher-education research; observational mixed-methods study at one public college. |
| [AVAILABLE SOURCES OR DATA] | List or paste the papers, notes, datasets, metadata, or repository records you may use. | Twelve articles, anonymized attendance records for 420 students, commute survey responses, and interview notes from 18 volunteers. |
| [METHODS AND CONSTRAINTS] | State your planned methods, sample limits, ethics requirements, tools, and deadline. | Descriptive and regression analysis with age, work hours, and course load considered; IRB-approved protocol; six-week deadline. |
| [OUTPUT OR PUBLICATION GOAL] | Describe the deliverable, target audience or journal, and required format. | A 3,000-word internal research report for student-success leadership, with methods appendix and citations. |
How to use
- Paste your actual question, source inventory, and methodological limits into the prompt.
- Generate the pack, then use each prompt with the relevant excerpts, records, or citations attached.
- Check every cited claim against the original paper or data and preserve a traceable note of where it came from.
- Follow up with: “Turn this evidence table into a claim map showing which claims are supported, mixed, or unsupported.”
Variations
Literature search
Use before beginning a reproducible literature search.
Develop a reproducible literature-search plan for [RESEARCH QUESTION] in [DATABASES], using [SCOPE LIMITS] and [INCLUSION EXCLUSION CRITERIA]. Output concept blocks, controlled-vocabulary suggestions to verify, Boolean strings for each database, date and language limits, duplicate-handling steps, and a screening log template. Identify likely terminology ambiguities and adjacent fields that may create false positives. Do not claim the search was run or provide invented citations. Ask up to 3 questions only if necessary inputs are missing.
Evidence synthesis
Use after you have extracted studies into notes or a table.
Synthesize these extracted study records: [STUDY RECORDS], to answer [RESEARCH QUESTION] for [INTENDED AUDIENCE]. Produce a structured narrative with scope, patterns, disagreement, study-quality limitations, applicability limits, and research gaps. Tag each conclusion with the studies or rows that support it, and distinguish association, mechanism, and causal evidence. Do not add sources, quotes, effect sizes, or certainty not present in the records. Ask up to 3 questions only if required context is missing.
Peer review critique
Use to prepare constructive review comments on a manuscript.
Review this manuscript or section: [MANUSCRIPT TEXT], against [REVIEW CRITERIA] and [JOURNAL OR VENUE]. Output prioritized comments under major concerns, minor concerns, reporting gaps, and questions for the authors. Tie each comment to a specific passage or missing methodological detail, and suggest a feasible revision where possible. Do not rewrite results, invent flaws, or make claims outside the submitted text. Keep tone professional and confidential. Ask up to 3 questions only if essential.
Tips
- Keep a citation-level evidence table before drafting prose; it prevents a polished synthesis from outrunning the underlying studies.
- Use AI-generated search strings as a starting point, then test them against known relevant papers and document every revision.
- Ask AI to label causal language separately from associative findings, especially when summarizing observational work.
- For qualitative work, have a domain researcher review proposed codes against raw excerpts and your reflexive memo.
FAQ
Can AI find all relevant studies for my review?
No. It can help design searches and screening criteria, but database execution, deduplication, and source verification remain necessary.
Can I trust AI-generated citations?
Do not treat them as reliable until you verify each item in the original database, publisher record, or reference manager.
Can AI analyze sensitive research data?
Only use systems approved for your data classification, consent terms, institutional policy, and ethics protocol.